Auto attentuator for input

On a sunny day (Thu, 08 Nov 2007 07:09:24 +1100) it happened Norm Lucas wrote in :

Sure, it is called 'auto ranging'. You could switch in attenuators or use some sort of AGC.

Yes. Peak detector, time constant (attack / decay) for reaction time. Question then is: What are you measuring, what is the min voltage, what is the max voltage, what is the dynamic behaviour?

If you don't care about the amplitude, maybe a log amp. would do the trick.

Al

It wouldn't need a micro, but to attempt a complete reply, I'd need to know the nature of the signal(s).

A crude system could have the Vu circuit from the LM3914/5/6/7/8/9/10 datasheet with the topish o/p decreasing a counter and a bottom o/p increasing the counter which alters some cmos switches that fiddlydum the gain.

Added bonus flashing lights

Martin

The input signals would be from any audio frequency source, such as one might encounter repairing audio gear. The PC software previously mentioned simply provides an oscilloscope-like display of the waveform.

Someone suggested an AGC, but my impression is that this would distort the input signal, including the amplitude. The response needs to be proportionate.

I already have a manaully activated attentuator and signal amp. The idea is to make it automatic, and in a single functional unit, so continual adjustment is not required with each different amplitude of input signal.

Any suggestions along these lines would be appreciated.

Norm Lucas

Quick question.

What dynamic range do you have with the source signal? might help with the discussion

Martin

10dB would suffice. Thanks for asking.

Norm

On a sunny day (Thu, 08 Nov 2007 13:20:53 +1100) it happened Norm Lucas wrote in :

You'd have to install Linux, the AGC works over much more then 10dB. AGC does not really distort the signal, it very slowly (about 30 seconds) increases gain until the peak amplitude is at some set level.

10dB is not even 4x.

I wouldn't bother with automating anything for this application. There are simple low parts count configurations for amplifiers with precision clamping. You can use that and if your displayed acquisition shows clipping then you would switch in a lesser gain via a wafer switch until it goes away. Automating the thing will never be more than a hunch, not to mention a pain in the butt, because of the random nature of the signals and the wide frequency range and mix.

To me that means it may be speech, music, noise, steady tone, or whatever. Is that what you are saying?

Proportionate to what? Any means of adjusting the signal will introduce distortion during periods of change. A misadjusted or poorly designed AGC can cause severe distortion, but a correctly operating one should have only "acceptable" distortion.

A good AGC/amp *could* do the job if the control attack and decay times were made manually variable to meet your changing needs. Mine was used to assure conformance to a telco tariff for signals at a network interface; it has a control range of over 40 dB.